Cosy Aromas, a store located in Ipswich, UK, has run Prospecting, Retargeting, and Cross-Sell ad campaigns with RetargetApp to get high ROAS and low cost per purchase.
Artificial Intelligence on Facebook and Google Ads: How does it work?
If you use the Internet every once in a while, you must have heard about machine learning and artificial intelligence (AI). The topic has been controversial for years — even the greatest of minds of our time cannot agree on it. Remember the discussion between Elon Must and Mark Zuckerberg? Here’s a quick recap.
What the two titans of modern-day technology couldn’t agree on was the superintelligence powered by machine learning. They argued about robots that may potentially outperform human intellect and then find us, people, redundant. However, even if such threat arises, it won’t happen in the near future. The human-like robots of today are actually very sweet.
AI is not only about robots. This technology is already used across industries from virtual assistants and self-driving cars to banking and yes, advertising. Since we are all here for the latter, let’s take a closer look at what AI advertising is, how the largest ad platforms use it and what’s in it for eCommerce store owners and digital marketers.
How do advertising algorithms work?
The working principle behind the targeted advertising algorithms is pretty straightforward. It’s all about the right match based on rich data. Ad platforms take the information about the product and targeted audiences on one hand, and about the people that are viewing web content on the other hand. The ad is shown if there’s a match of interests, buying intent, demographics, etc.
However, there are many advertisers on ad platforms, and often they compete for the same users. To reach their potential customer, the advertisers enter an auction. With ad auctions, it’s not only the bid size that determines the winner.
Here are some of the factors that determine who wins the ad auction:
- Budget: the higher your budget is, the more chances your ad has to be shown
- Relevance: the most relevant ad will be shown to a person
- Interactions: it matters how people interact with the ad, e.g. tap on it, share it, hide it, report it, etc.
- Compliance with the platform’s rules: if there is even a potential breach of the ad platform’s rules (e.g. Facebook Advertising Policies), the chances to win the auction decrease significantly.
How is AI used in advertising?
First, let’s see what artificial intelligence and machine learning mean.
Artificial intelligence (AI) is a broad concept that describes the machine’s ability to perform tasks and solve problems creatively — just as humans do. It is also about designing machines that can think, reason, and behave like people.
Machine learning (ML) is a way to apply AI to solve problems. Its main principle is that once a machine accesses data, it can automatically learn how to find the solution without being programmed for each specific task.
That’s all well and good, but what does it have to do with advertisement? Artificial intelligence and machine learning process complex information in large volumes to come up with smart advertising solutions. Ad platforms and advertisers adopted this technology to:
- Optimize the process of ad delivery
- Take the workload off advertisers
- Minimize chances of human error
- Find the best-performing ad visuals and copy.
It may seem that AI is the new kid on the advertising block. To be fair, it became a buzzword only a couple of years ago. However, AI made its way to the world of marketing in 2014 when programmatic advertising — an automated process of buying ads — became a thing, quite a big thing, actually.
This technology enabled brands to automatically buy ad slots (spaces where ads are displayed) from websites online. Ad delivery to relevant audiences became faster, more efficient, and cheaper since all the manual labor was removed from the ad buying equation.
The future of AI in programmatic advertising
Modern-day marketers may see the advertising methods from the pre-programmatic days as archaic and grossly inefficient. With artificial intelligence under the hood, the speed and performance of today’s advertising are by far more superior to that of several decades ago. This means that in the future, the use of AI in programmatic advertising will become even more indispensable.
The aspects that AI will keep handling in programmatic advertising in the future are:
- Ad personalization by processing vast amounts of data about a user to choose the most relevant ads for them
- Ad placement by analyzing the content on the website and the ad’s copy to understand which ad will suit it best
- Ad spend by automatically analyzing such details as website ranking, relevance, etc., and deciding on the optimal bid for the ad slot to prevent overspending
- Ad analytics by providing predictive analysis based on similarities between ad viewers and existing customers.
The flip side of programmatic advertisement
Programmatic advertising comes with a price. To see meaningful results, businesses need to have large budgets. The ads themselves cost big bucks, however, brands must also be prepared to spend money without immediate return to gather data and find the right audiences and placements.
After reading all this, you may be left with the question: how do I use AI to promote my eCommerce store? We’ve got the answer that you may like: the largest ad platforms, Facebook and Google, are on your side. Both have implemented AI into their platforms to give small and medium online store owners the chance to leverage this technology for their marketing.
How do Facebook and Google decide which ads to show?
Facebook Ads and Google Ads were quick enough to spot the potential of machine learning and invest in this technology — and it was a great idea. Here’s how the two platforms use machine learning, and what’s in it for you.
Facebook ML algorithms
Facebook Ads bring value to both advertisers and Facebook users. One of the ways to bring value is by optimizing ad delivery, so users see the right ads and marketers have more conversions and high return on the ad spend (ROAS). After seeing that AI and machine learning started outpacing the marketers’ manual work, Facebook started making the AI-based technology accessible not only to large ad agencies with their army of marketers, but also to small business owners. Here is how Facebook Ads empower eCommerce stores and digital marketers with the machine learning advertising tools:
Power 5 is an AI-powered framework that drives the best return on investment (ROI) for advertisers. This framework includes 5 core tactics:
- Auto advanced matching to reach more relevant audiences and increase conversions by giving Facebook hashed (converted into lines of characters) customer details and Pixel events.
- Simplified account structure to let advertising algorithms identify the best-performing creatives and platforms to optimize campaigns in real time.
- Campaign budget optimization to spend the money only on the best-performing campaigns by setting one budget that will be spent on different ad sets.
- Automatic placements to automatically display ads to the relevant audiences across different placements such as Facebook, Instagram, Messenger, WhatsApp, etc.
- Dynamic ads to deliver targeted ads to people based on the content they viewed in the web store and the actions they took.
Delivering ads to people is not the only thing that matters in digital advertising. Tracking performance is crucial to understand if the strategy is working, and what to tweak in case it isn’t. That is why Facebook has introduced data-driven attribution (DDA) model powered by machine learning.
DDA is an attribution model that measures the progressive results you receive with Facebook Ads. It demonstrates how people’s actions on Facebook result in a conversion. Facebook calls such actions “touchpoints”. They include impressions, clicks, and web store visits before the actual conversion.
To make sure that such attribution model is accurate, it is checked for bias. All results received under the DDA are compared with the Facebook conversion lift studies that didn’t use this model. This method ensures that there are no over- or underestimations of the results.
Google ML Algorithms
Of course, Facebook is not the only ad platform that flexes AI muscles. Google, another giant in the advertising market, has been honing the machine learning powered tools for some time now.
The platform offers advertisers AI technology to reach the following objectives:
Delivering highly relevant ads on YouTube
Long gone are the days when YouTube was only a place to watch funny videos. Now, it is a powerful platform that helps people make a purchase decision. For example, every other car buyer watches a video review before deciding on the car they want. Furthermore, every other millennial searches for cooking tips on YouTube before doing grocery shopping.
The task of an advertiser is to catch the person’s attention at the right moment. Machine learning helps them do just that with the Maximize Lift bidding. It delivers ads to people who are potentially interested in your product or brand.
Local campaigns to bring people to the physical store
While eCommerce is booming in the world of social distancing, many people still prefer buying items in a brick-and-mortar store. Thus, 80% of buyers will choose to go to a shop when they need a product right away. The number of “Near me” searches has increased by 3 times.
That is why Google gives store owners opportunities to bring potential customers to the physical shop with the Store visits ad objective. The business provides its address and ad creative, and Google delivers such ads to people looking for related products in their location.
How does machine learning improve ad delivery?
When the ad campaign is running, the stakes are high and the time is limited. In a matter of seconds, Facebook needs to match an ad with the people that are most likely to act on it. Can you imagine how tedious it would be to manually connect users and ads, given that Facebook has over 2 billion users? It would be also nearly impossible for developers to program the platform to make the optimal decision each time.
With machine learning, Facebook delivers relevant AI-powered ads fast and effectively. Facebook ad algorithms automatically analyze such information as the business objectives of an advertiser and the users’ behavior to understand how likely a person is to take the target action — visiting a website, signing up for an event, or making a purchase.
In turn, Google Ads offers responsive ads to automatically meet online users’ needs faster without extra effort. All an advertiser needs to do is come up with 4 descriptions and 15 headlines. Machine learning will then combine them depending on the search query people make. The efficiency increase is quite impressive: businesses that run responsive search ads get up to 15% more clicks.
Benefits of using AI in advertisingAd automation
Optimizing Ad Spend with AI
ML helps businesses predict their revenues and provide tips on optimizing the campaigns. This is possible due to algorithms of machine learning to track how people respond to your ads and the ads of your competitors. For example, there is a misconception that only ads with higher bids get delivered. In fact, very often the ads with lower bids win the auction if the algorithms see that an ad is more relevant to the person. This offers equal opportunities for all players on the market. Megabrands with megabudgets and a small online store have the same chances of showing ads.As you have already seen above, Facebook and Google ad platforms offer truly great AI capabilities for ad management. However, it does take some time and effort to explore these opportunities and apply them in your day-to-day marketing practices. The good news is that there are solutions that help businesses put AI to work in a matter of clicks.
The reason artificial intelligence helps boost ROI lies in the fact that AI makes advertising data-driven. Here’s why artificial intelligence impacts your ROI:
- Targeting people that have high purchase intent with highly personalized ads including people similar to your clients and/or website visitors
- Processing the data on the previous performance together with the markets situations to give you a better understanding of which campaigns and audiences work best for your goals
- Cross-selling and upselling the products to people based on their purchase history by tracking people’s activity on your website.
Artificial Intelligence for ad creation
AI can also help you create and adjust content. Different ad solutions find their own way to use machine learning for this purpose. Some provide tools for analyzing the potential performance of ad creatives and copy. Others take marketing further than ad delivery and help optimize email marketing content as well.
However, there may be a risk of relying too heavily on AI and publish content without double-checking it. This may cause some awkward situations where your text doesn’t make much sense. Even though machine learning is a powerful tool, it still has to go a long way before it can substitute people. When you do use AI for content creation, make sure that your copy meets your goals and is easy to understand.
Targeting the perfect audiences
To make targeting as precise as possible, Facebook’s artificial intelligence advertising algorithms take into account not only what a user does on Facebook, but also how they interact with the business off the social platform. For example, ML tracks person behavior in the web store: if and when they visited the site, which products they viewed, if they added anything to the cart, etc.
Based on this information, advertisers use Retargeting to motivate the website visitors to make the purchase. To be able to track this information for Retargeting, businesses owners need to verify their domain on Facebook and install Pixel, a piece of code that helps track the efficiency of Facebook intelligent ads and how people interact with the content on your website.
How to get started with AI in advertising with RetargetApp
RetargetApp enables business owners to set up their first campaign fast and stress-free. Plus, the app automates Retargeting and other campaigns, which allows merchants to check on them every once in a while.
RetargetApp helps eCommerce store owners and digital marketers automatically set up a campaign in one place faster than on Facebook and Google to receive, on average, 600% ROAS.
In the app, you can set up Retargeting with Facebook Dynamic Ads to target your visitors based on how they interacted with your products. Also, you can effortlessly run a Prospecting campaign that targets people who are similar to your existing clientele. As a result, marketers spend less time on creating and managing campaigns manually while receiving higher results thanks to machine learning algorithms.
As for Google, RetargetApp lets you run the Smart Shopping ads to increase your ROAS by reaching out only to those people who are most likely to convert.
It will take no more than 15 minutes to set up the app and a couple more minutes to create an AI-powered campaign. Learn more about ads automation here.
Artificial intelligence and machine learning are nowhere near threatening digital business. On the contrary, they are making people’s lives a lot easier. Digital advertising is one of the things that AI makes more effective. That is why Facebook and Google have invested so much money and effort into powering their ad platforms with machine learning technology.
It was totally worth it. Now advertisers can run highly relevant ads to grow ROAS and sales without constantly tracking the ad performance and fine-tuning the settings of their campaigns every day. Ad solutions outside of Facebook and Google took this technology and offered even more effortless automation that enables businesses owners and marketing managers to spend minutes a day — or less — on setting up and optimizing ads.
It may be difficult to wrap your head around AI in advertising and how to make sure your ads benefit from it. The good news is that you’re not alone in it. RetargetApp helps you ensure that machine learning algorithms by Facebook and Google drive the best results for your campaigns.